Multi machine barebox builds have to pass information on which
board we are running on via boarddata. Usually this will be a
pointer to a device tree. Some boards might not have a device
tree available though because they are either not ported over
to device tree yet, or are running in some limited first state
environment which does not offer enough space for a device
tree. For these cases this patch adds a mechanism to embed a
machine number into a struct type along with a magic number.
This makes it possible to check for a specific machine
later during regular runtime.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This adds a barrier after setup_c(). This is necessary because otherwise
some global variable assignments may be reordered by the compiler to be
executed before setup_c which cannot work.
This was observed when doing other unrelated changes to the start function,
it seems in current mainline state the compiler does not actually reorder
the code.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Those should only be used internally. All users should rather
use the streaming DMA API, which does proper cache maintenance.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
As a lot drivers currently rely on the 1:1 virt->phys mapping on ARM
we define DMA_ADDRESS_BROKEN to mark them. In order to use them on
other architectures with a different mapping they need proper fixing.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Sometimes it's useful to test if a memory operation works or
aborts. This adds data_abort_mask() to ignore data aborts and
data_abort_unmask() to enable them again. This is used in the
next step for the 'md' command so that illegal addresses just
show 'xxxxxxxx' instead of crashing the system.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
So users can pass in device memory pointers without provoking
warnings.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The relocated_to_adr does not clear the .bss section.
It is done by the setup_c routine.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
These comments are wrong. Anyway, they do not carry useful
information. Delete.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Add the pxa3xx architecture, for pxa300, pxa310 and pxa320 SoCs.
This includes :
- the registers
- the cpu type
- the reset source handling
- a minimal set of clocks
- pin control definitions
In this last case, the big mfp-pxaxxx.h files were dropped, forcing
board developers to use MFP_LPM_* macros cunningly.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This file originates in Linux. Linux has it under include/linux/
directory since commit dccd2304cc90.
Let's move it to the same place as well in barebox.
This commit was generated by the following commands:
find -name '*.[chS]' | xargs sed -i -e 's:<sizes.h>:<linux/sizes.h>:'
git mv include/sizes.h include/linux/
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Now that we have printf support in the PBL we can use it right
after setup_c(). Add some debug messages to the early PBL code
to make it more clear what is happening there.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
pr_debug can now be used right after setup_c(), so add some debug
messages to the early startup code to make it a bit more clear what
is happening there.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The corresponding code doesn't use the lr register (neither explicitly
nor implicitly by the bl instruction), so there is no gain in using r2
here.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
A part of the existing comments was incomplete or missleading.
Adding the register name to mcr/mrc instructions helps finding the
corresponding documentation in the manuals.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Architecturally the cache contents are undefined so it might well
contain stale data at reset. So better be save than sorry.
I verifyed that the added instructions are defined for both, ARMv6 and
ARMv7, using the ARM Architecture Reference Manual, ARMv7-A and ARMv7-R
edition (ARM DDI 0406C.c). For the already existing mcr instruction see
the newly added comment.
This patch also unifies handling of ARMv6 and ARMv7, the isb instruction
can also be done on the latter via mcr which simplifies the code a bit.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
CPU_ARM946E is selected by boards, so letting it depend on !MMU
leads to broken dependencies. Let MMU depend on !CPU_ARM946E instead.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
ARM946E-S is used in the DIGIC family SoCs.
ARM946E-S core supports ARMv5TE and has Cache & MPU.
Linux kernel uses ARMv4 MPU cache routines for ARM946E-S core.
E.g. see linux.git/boot/compressed/head.S:
.word 0x41009400 @ ARM94x
.word 0xff00ff00
W(b) __armv4_mpu_cache_on
W(b) __armv4_mpu_cache_off
W(b) __armv4_mpu_cache_flush
So select CPU_32v4T for ARM946E-S despite of ARMv5TE support.
Signed-off-by: Antony Pavlov <antonynpavlov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Device trees can be passed by primary boot loader or appended to
barebox binary. Unfortunately, before probing devices from such
a device tree, no fixup can be applied.
Add a call to of_fix_tree() right before probing devices to catch
some very early fixups.
Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
We need to flush out the ttb in order to make the
changes observable to the page walker.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This increases the malloc to half of the available memory in a
bank. This helps with some usecases requiring a lot of memory.
The other half is still available as scratch area and for
putting the kernel binary.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
We used to relocate the executable to the current address. This does
not work when the executable runs from a readonly location like for
example NOR Flash. Test if we run from inside the available memory and
if we do, relocate to the current address as before. Otherwise copy
the executable to the start of memory and relocate to that address.
While at it add some comments to the code.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The FSF address has changed; The FSF site says that
address is
Free Software Foundation
51 Franklin Street, Fifth Floor
Boston, MA 02110-1301
USA
(see http://www.fsf.org/about/contact/)
Instead of updating it each time the address changes,
just drop it completely treewide.
Signed-off-by: Antony Pavlov <antonynpavlov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
I assume I am the only person knowing that barebox is able to
merge devicetrees. This feature seems broken for a while now since
trying to merge devicetress results in:
unflatten: too many end nodes
Remove this feature to save the complexity.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This patch does probably too much, but it's hard (and very
cumbersome/time consuming) to break it out. What is does is this:
* each command has one short description, e.g. "list MUX configuration"
* made sure the short descriptions start lowercase
* each command has one usage. That string contains just the
options, e.g. "[-npn]". It's not part of the long help text.
* that is, it doesn't say "[OPTIONS]" anymore, every usable option
is listed by character in this (short) option string (the long
description is in the long help text, as before)
* help texts have been reworked, to make them
- sometimes smaller
- sometimes describe the options better
- more often present themselves in a nicer format
* all long help texts are now created with BUSYBOX_CMD_HELP_
macros, no more 'static const __maybe_unused char cmd_foobar_help[]'
* made sure the long help texts starts uppercase
* because cmdtp->name and cmdtp->opts together provide the new usage,
all "Usage: foobar" texts have been removed from the long help texts
* BUSYBOX_CMD_HELP_TEXT() provides the trailing newline by itself, this
is nicer in the source code
* BUSYBOX_CMD_HELP_OPT() provides the trailing newline by itself
* made sure no line gets longer than 77 characters
* delibertely renamed cmdtp->usage, so that we can get compile-time
errors (e.g. in out-of-tree modules that use register_command()
* the 'help' command can now always emit the usage, even without
compiled long help texts
* 'help -v' gives a list of commands with their short description, this
is similar like the old "help" command before my patchset
* 'help -a' gives out help of all commands
Signed-off-by: Holger Schurig <holgerschurig@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The old output of "help" was just producing a long list, that usually
scrolled of the screen (even on a X11 terminal). This list is more
compact, and also sorted by groups.
The old output format (plus grouping) is now available with 'help -v'.
Example:
Information commands:
?, devinfo, help, iomem, meminfo, version
Boot commands:
boot, bootm, go, loadb, loads, loadx, loady, saves, uimage
...
Signed-off-by: Holger Schurig <holgerschurig@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Mostly to make it clear that boarddata needs to be
something we can dereference.
As this is a pretty invasive change, use the opportunity
to make the signature 64bit safe.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
For memory reaching the end of the address space
phys + bank->size overflows to 0. Fix this by right shifting
phys and bank->size before adding them.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Tested-by: Sean Cross <xobs@kosagi.com>
- If we have no memory registered in mmu_init() it's a critical bug.
panic in this case.
- If we do not have a ttb when dma_alloc_coherent or remap_range is
called it's also a critical bug. Panic in this case.
- if find_pte is called with an address outside our memory banks dump
the memory banks and the address to give more clue what went wrong.
Also add some hints what might went wrong to the code.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
BUG() uses printf which is not available in pbl, so do not use it here.
This becomes necessary when multiple CPU architectures are compiled in.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The cache-v7 code uses assembler instructions which do not
exist on before v7, so explicitely pass armv7-a to this file
to make the compiler happy.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The multi image startup process used to have three binaries involved:
- The lowlevel board code to initialize SDRAM
- the uncompressor
- the regular (compressed) barebox binary
Drop the uncompressor and put the uncompress code into the lowlevel
board code binary. This makes the startup process easier.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>